Journal article on regression in model space

The article "Modelling of Parametrized Processes via Regression in the Model Space of Neural Networks" by W. Aswolinskiy, F. Reinhart and J.Steil
has just been published in Neurocomputing and is available free of charge until 20.Oct from here. The paper pursues a meta-modelling approach, where the parametrization of a particluar network is learned by another network on the next higher level,
that is in the model space.